bugfix

app.py

@@ -15,7 +15,7 @@ model = BlipForConditionalGeneration.from_pretrained(
 
 def inference(raw_image, model_n, strategy):
     if model_n == 'Image Captioning':
-        input = processor(raw_image).to(device, torch.float16)
+        input = processor(raw_image, return_tensors="pt").to(device, torch.float16)
         with torch.no_grad():
             if strategy == "Beam search":
                 config = GenerationConfig(
@@ -38,7 +38,8 @@ def inference(raw_image, model_n, strategy):
             return 'caption: '+caption
 
 
-inputs = [gr.inputs.Image(type='pil'), gr.inputs.Radio(choices=['Image Captioning'], type="value", default="Image Captioning", label="Task"), gr.inputs.Radio(choices=['Beam search', 'Nucleus sampling'], type="value", default="Nucleus sampling", label="Caption Decoding Strategy")]
+inputs = [gr.inputs.Image(type='pil'), gr.inputs.Radio(choices=['Image Captioning'], type="value", default="Image Captioning", label="Task"), gr.inputs.Radio(
+    choices=['Beam search', 'Nucleus sampling'], type="value", default="Nucleus sampling", label="Caption Decoding Strategy")]
 outputs = gr.outputs.Textbox(label="Output")
 
 title = "BLIP"